Use of vibration signal to estimate instantaneous angular frequency under strong nonstationary regimes

Instantaneous angular frequency (IAF) is an inherent characteristic associated with mechanical dynamics and contains adequate diagnostic information about rotating machinery, especially under speed fluctuation conditions. However, not all machines have the facility to install tachometers or encoders for tracking the instantaneous angular speed (IAS). An alternate solution is using vibration measurements and recovering the missing speed by exploring the acceleration signal. Thus, accurate IAF estimation is crucial to the consequent fault diagnosis. To date, multiple methods have been developed to address this issue, but most present poor results when dealing with the significant variations of IAS, which are common in practical applications. To overcome this deficiency, a new method consisting of time–frequency ridge extraction and an improved multi-order probabilistic approach is proposed in the paper. The described method fully exploits the information within the vibration response while promising robustness to the variations of the running regimes. The proposed method is demonstrated by assessing its performance on three benchmark data sets and two experimental signals collected from our laboratory and comparing them with five advanced IAF estimation methods.